Update app.py

app.py

@@ -424,6 +424,7 @@ def generate_text_zu_doc(file, prompt, k, rag_option, chatbot, history, db):
 #mit oder ohne RAG möglich
 def generate_text (prompt, chatbot, history, rag_option, model_option, openai_api_key, db, k=3, top_p=0.6, temperature=0.5, max_new_tokens=4048, max_context_length_tokens=2048, repetition_penalty=1.3, top_k=35):
     global splittet
+    hugchat=False
     suche_im_Netz="Antwort der KI ..."
     print("Text pur..............................")
     if (openai_api_key == "" or openai_api_key == "sk-"):
@@ -465,6 +466,7 @@ def generate_text (prompt, chatbot, history, rag_option, model_option, openai_ap
             #llm = HuggingFaceTextGenInference( inference_server_url="http://localhost:8010/", max_new_tokens=max_new_tokens,top_k=10,top_p=top_p,typical_p=0.95,temperature=temperature,repetition_penalty=repetition_penalty,)
             #llm via HuggingChat
             llm = hugchat.ChatBot(cookies=cookies.get_dict())
+            hugchat=True   #da dieses Model in llm_chain bzw reag_chain anderes verarbeitet wird
             
             print("HF")
             #Prompt an history anhängen und einen Text daraus machen
@@ -473,16 +475,11 @@ def generate_text (prompt, chatbot, history, rag_option, model_option, openai_ap
         #zusätzliche Dokumenten Splits aus DB zum Prompt hinzufügen (aus VektorDB - Chroma oder Mongo DB)
         if (rag_option == "An"):
             print("LLM aufrufen mit RAG: ...........")
-            result = rag_chain(llm, history_text_und_prompt, db)
-        #elif (rag_option == "MongoDB"):
-            #splits = document_loading_splitting()
-            #document_storage_mongodb(splits)
-            #db = document_retrieval_mongodb(llm, history_text_und_prompt)
-            #result = rag_chain(llm, history_text_und_prompt, db)
+            result = rag_chain(llm, history_text_und_prompt, db, hugchat)
         else:
             #splittet = False
             print("LLM aufrufen ohne RAG: ...........")
-            resulti = llm_chain(llm, history_text_und_prompt)
+            resulti = llm_chain(llm, history_text_und_prompt, hugchat)
             result = resulti.strip()
             """
             #Alternativ mit API_URL - aber das model braucht 93 B Space!!!